Speaker verification using relaxation labeling

Tuan Pham, Dat Tran, Michael Wagner

Research output: Contribution to conference (non-published works)Paperpeer-review

Abstract

A relaxation algorithm for speaker verification is presented in this paper. This verification system uses a nonlinear probabilistic framework to paralelly and iteratively update the labelings of the input speech feature vectors until no further improvements are found. The implementation is tested on a commercial speech corpus TI46. The equal error rates obtained from the proposed approach are significantly more favorable than those from the conventional VQ-based method.

Original languageEnglish
Pages29-32
Number of pages4
Publication statusPublished - 2020
EventWorkshop on Speaker Recognition and its Commercial and Forensic Applications, RLA2C 1998 - Avignon, France
Duration: 20 Apr 199823 Apr 1998

Conference

ConferenceWorkshop on Speaker Recognition and its Commercial and Forensic Applications, RLA2C 1998
CountryFrance
CityAvignon
Period20/04/9823/04/98

Fingerprint

Dive into the research topics of 'Speaker verification using relaxation labeling'. Together they form a unique fingerprint.

Cite this